One of the core symptoms of many anxiety disorders, especially Post-Traumatic Stress Disorder (PTSD), is an inability for fear safe in situations where healthy individuals do feel safe. Thus, animal models of fear conditioning and fear inhibition offer useful tools to determine how these learned fears are diminished or inhibited. We have developed a new paradigm in rodents referred to as AX?, where cues A and X in compound signal an aversive event and cues B and X in compound signal no aversive event (safety). In a critical subsequent transfer test trial, presentation of A and B together (AB) results in a reduced fear response compared with the response to A. Other tests have shown this is bona fide conditioning inhibition and not due to external inhibition. We have now found this to be true in humans and rhesus monkeys where we see transfer of inhibition on AB test trials, in contrast to prior failures to see transfer in humans using the typical conditioned inhibition paradigm. Most importantly, in three independent groups of patients with PTSD we see some discrimination between AX and BX but no transfer on the critical AB test trial, thus detecting a core symptom in PTSD. The R21 phase of this application is to modify our current AX? paradigm into a "working memory" test, which will allow the same monkeys to be tested repeatedly in this new paradigm using sets of pictures as stimuli instead of lights and tones. We will then evaluate in adult monkeys that have sustained neurotoxic lesions of orbital frontal areas 14/25 vs.11/13 vs. 12 in safety signal learning and expression. As a positive control, we will also test these monkeys in reversal learning and reinforce devaluation that is known to be compromised by damage to one or more of these lesions. If successful, the R33 phase will begin to evaluate the development of safety signal learning from year 1 to year 3, a time period when pronounced developmental changes occur in these orbital frontal areas. We believe a "working memory" version of this measure of safety signal learning in which the same animal can be tested repeatedly will provide a major new paradigm to study safety signal learning in psychiatric disorders and to eventually lead to new and better treatments for people with anxiety disorders. This project is clinically relevant because: (1) many emotional disorders in humans, such as anxiety, phobias and post-traumatic stress disorders, are characterized by a resistance to extinguish learned emotional reactions to anxiogenic stimuli or contextual information associated with these anxiogenic stimuli, (2) learned fear in early infancy has strong resistance to extinction that yield anxiety disorders later in life and (3) anxiety disorders have also been reported in several developmental neuropsychiatric disorders, such as autism and schizophrenia, as well as following pediatric traumatic brain injury and early life stress.